<!-- Generated by scripts/generate-plugin-artifacts.py; do not edit. -->

Codex wrapper for /benford

Follow the OloLand command instructions below when the user asks for /benford or the equivalent workflow in Codex.

Benford's Law

Benford's Law states that in many naturally-occurring datasets, the leading digit follows a logarithmic distribution: ~30% start with 1, ~17% with 2, etc. Significant deviation from this distribution is a classic forensic indicator of fabricated or selectively-entered transactions.

Usage

/benford <deal_id>

Arguments

• deal_id (required) — The deal whose GL transactions to test. Requires uploaded GL export with at least 1,000 line items for statistical validity.

Execution

1. If the user supplied a transaction list or GL file, parse it into transaction dictionaries and call run_benford(transactions). If the user supplied only deal_id, call analyze_forensic_qoe(deal_id) and extract the Benford section from the full battery. Do not pass a primitives argument.
2. The engine pulls all GL transactions from the deal's data room (or from forensic_extraction_service if extraction has been run).
3. Computes observed first-digit distribution and compares to Benford expected via χ² goodness-of-fit and Mean Absolute Deviation (MAD).
4. Flags account categories with the most significant deviations.

Interpretation

• MAD < 0.012 — close conformity (green)
• 0.012 ≤ MAD < 0.015 — acceptable conformity (yellow)
• MAD ≥ 0.015 — non-conformity (red — investigate the high-deviation accounts)

Output

For each account category that deviates significantly from Benford expected:

• Observed vs. expected first-digit distribution (table + chart)
• χ² statistic and p-value
• Top transactions contributing to the deviation
• Suggested follow-up: which specific journal entries to examine

Why this matters

Benford's Law is the deterministic statistical test most commonly cited in fraud forensics literature. It cannot be evaded by altering totals — it operates on the digit-level distribution of every transaction. Pre-LOI screening with Benford catches the deals where journal entries have been manipulated before you commit to fieldwork.

Example

/benford deal_acme_2026